B. — CHEMISTRY. 



55 



' Contaminated with silicon. 



'' Known to be imiJure. 



With the exception of silver and gold, these metals were the purest 

 oBlainable in commerce. 



Two facts are evident from the consideration of this table : (a) the 

 remarkable accuracy of Callendar's formula connecting the Tempera- 

 ture Centigrade with the change of resistance of a pure platinum wire ; 

 (b) the accuracy of Callendar and Griffiths' determination of the boiling- 

 point of sulphur. Although the platinum resistance pyrometer had at 

 this time only been compared with the air thermometer up to 600° C, 

 it will be noted that the exterpolation from 600° to nearly 1,100 was 

 justified. 



I cannot leave the subject of high-temperature measurements with- 

 out referring to the specially valua,ble work of Burgess, and also to 

 Eza Griffiths' book on high-temperature measurements, which contains 

 an excellent summary of the present state of our knowledge of this 

 important subject. 



During the period that the above work on non-ferrous alloys was 

 being done, great progress was being made in the study of iron and 

 steel by Osmond and Le Chatelier. In 1890 the Institute of Mechanical 

 Enguieers, not apparently without considerable misgivings on the part 

 of some of its members, formed an Alloys Eesearch Committee. This 

 Committee invited Professor (afterwards Sir Wilham) Eoberts-Austen 

 to undertake research work for them. The results of his investigations 

 are contained in a series of five valuable reports, extending from 1891 

 to 1899, published in the Journal of the Institute. The first report 

 eontained a description of an improved form of the Le Chatelier record- 

 ing pyi-ometer, and the instrument has since proved a powerful weapon 

 of research. In the second report, issued in 1893, the effects on the 

 properties of copper of small quantities of arsenic, bismuth, and 

 antimony were discussed. Whilst some engineers advocated, others 

 as strongly controverted, the beneficial results of small tjuantities of 



